Revealing phenylium, phenonium, yinylenephenonium, and benzenium ions in solution

Article abstract of DOI:10.1002/chem.200701043

The 4-methoxyphenylium ion has been generated in the triplet state ( ³An⁺) by photolysis of 4-chloroanisole in polar media and detected by flash photolysis (λ_max = 400nm). This is the first detection of a phenylium ion in solution by flash photolysis and the assignment is supported by time-dependent density functional theory (TD-DFT) calculations. In neat solvents, the cation was reduced to anisole, a process initiated by electron transfer from the starting compound (³An⁺ + AnCl→An^?+AnCf^?+, with the radical cation detected at 470 nm, then An^?→ AnH). Addition of π nucleophiles to the ³An⁺ cation offers a novel access to a number of other cationic intermediates under mild, nonacidic conditions. Two intermediates are successively formed with alkenes, a diradical cation and the phenonium ion, which are detected at 440 and 320 nm, respectively, by flash photolysis and are in accordance with calculations. Allylanisoles or β-alkoxyalkylanisoles are the end products, with a small amount of α-alkoxyalkylanisoles that arises from a Wagner-Meerwein rearrangement to form benzyl cations. Further intermediates that have been predicted and detected are the phenylvinylium ion, possibly in equilibrium with the vinylenephenonium ion, with 1-hexyne (λ_max = 340 nm) and the benzenium ion with benzene (λ_max = 380nm). The final products were anisylhexyne and methoxybiphenyl (an analogous product and intermediate were detected with thiophene).

Full text of DOI:10.1002/chem.200701043

FULL PAPER
DOI: 10.1002/chem.200701043
Revealing Phenylium, Phenonium, Vinylenephenonium, and Benzenium Ions
in Solution
Ilse Manet,[a] Sandra Monti,*[a] Gottfried Grabner,[b] Stefano Protti,[c] Daniele Dondi,[c]
Valentina Dichiarante,[c] Maurizio Fagnoni,[c] and Angelo Albini*[c]
3
Abstract: The 4-methoxyphenylium ion
has been generated in the triplet state
(3An+) by photolysis of 4-chloroanisole
in polar media and detected by flash
photolysis (lmax =400 nm). This is the
first detection of a phenylium ion in so-
lution by flash photolysis and the as-
signment is supported by time-depen-
dent density functional theory (TD-
DFT) calculations. In neat solvents, the
cation was reduced to anisole, a pro-
cess initiated by electron transfer from
the An+ cation offers a novel access
to a number of other cationic inter-
mediates under mild, nonacidic condi-
tions. Two intermediates are successive-
ly formed with alkenes, a diradical
cation and the phenonium ion, which
are detected at 440 and 320 nm, respec-
tively, by flash photolysis and are in ac-
cordance with calculations. Allylani-
soles or b-alkoxyalkylanisoles are the
end products, with a small amount of
a-alkoxyalkylanisoles that arises from
a Wagner–Meerwein rearrangement to
form benzyl cations. Further intermedi-
ates that have been predicted and de-
tected are the phenylvinylium ion, pos-
sibly in equilibrium with the vinylene-
phenonium ion, with 1-hexyne (lmax
=
340 nm) and the benzenium ion with
benzene (lmax =380 nm). The final
products were anisylhexyne and me-
thoxybiphenyl (an analogous product
and intermediate were detected with
thiophene).
Keywords: aryl cations
· carbo-
the starting compound
(3An+ +
AHCTREUNG
AnCl!An +AnCl +, with the radical
C
C
tions · photochemistry · reaction
mechanisms
C
cation detected at 470 nm, then An !
AnH). Addition of p nucleophiles to
Introduction
Of the conceivable intermediates in organic chemistry, diva-
lent carbocations, and in particular, phenyl cations (phenyl-
ium ions) are among the less well-known.[1] Two reasons ex-
plain the limited data available. The first is the scant syn-
thetic application of this intermediate. Phenyl cations are
known intermediates in some thermal or photochemical de-
composition reactions of phenyldiazonium salts,[2] and in
some cases, phenyl triflates.[3] The reactions occurring are
mostly solvolysis processes that are moderately appealing
from a synthetic point of view. This has led to the belief that
the excessive reactivity of phenyl cations makes them unse-
lective and of little preparative interest. On the contrary, the
majority of synthetically useful reactions of diazonium salts
(e.g., the Sandmeyer, Meerwein, and Gomberg reactions)
do not involve the phenyl cation but the phenyl radical.[4]
Some indications of a more varied chemistry were obtained
when the cation was generated by the decay of tritioben-
zene,[5] but this method is hardly suitable for synthesis.
[a] Dr. I. Manet, Dr. S. Monti
Istituto per la Sintesi Organica e la Fotoreattività-ISOF
CNR Area della Ricerca, Via P. Gobetti 101
40129 Bologna (Italy)
Fax : (+39)051-639-9844
E-mail: monti@isof.cnr.it
[b] Prof. G. Grabner
Max F. Perutz Laboratories, University of Vienna
Campus Vienna Biocenter 5, 1030 Wien (Austria)
[c] Dr. S. Protti, Dr. D. Dondi, V. Dichiarante, Dr. M. Fagnoni,
Prof. A. Albini
Dipartimento di Chimica Organica, Università di Pavia
Via Taramelli 10, 27100 Pavia (Italy)
Fax : (+34)038-298-7323
E-mail: angelo.albini@unipv.it
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author. It contains meth-
ods of irradiation, characterization of photoproducts, details of flash
photolysis measurements, optimized geometries listed in Cartesian
format, energies calculated (B3LYP 6-31G(d)) for 4-chloroanisoles
and adducts in the gas phase and in MeCN, and calculated spectra of
the cations.
The second reason is that mechanistic studies of the
phenyl cation in solution have been hampered by the fact
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1029

that this species appears to have no clear-cut UV absorp-
tion[6] and thus flash photolysis, which has been instrumental
in the study of other intermediates, such as carbenes and ni-
trenes, has been of limited value. Only a secondary inter-
mediate, the benzenium ion (cyclohexadienylium cation),
which results from the addition of the phenylium ion to ben-
zene, has been detected in solution.[7] Also, evidence for the
formation of the phenyl cation in matrices is scarce. EPR
spectra of the phenyl cation have been obtained by photoly-
sis of some phenyl diazonium ions in EtOH at 77 K,[8] and
only recently, the IR spectrum of the phenyl cation was ob-
tained by photolysis of phenyl iodide in an argon matrix at
4 K in which the phenyl radical is formed and gives the
cation by ensuing electron transfer.[9]
Results and Discussion
Photoproducts in various solvents: 4-Chlorophenol has pre-
viously been studied in detail,[14a,b,16] however, the studies
were affected by the protolytic equilibrium of the OH
group. Therefore, this work was carried out by using 4-chlor-
oanisole (1, see Scheme 2).
À
Recently, a class of reactions has emerged in which C C
bond formation has been achieved by the photolysis of elec-
tron-donating substituted aryl halides[10] or esters[11] in the
presence of carbon nucleophiles, for example, alkenes, al-
kynes, or (hetero)aromatics, with little competing solvolysis.
These reactions have been rationalized in terms of trapping
of the triplet phenylium ion (I, Scheme 1), which generates
Scheme 2. Photoproducts of 4-chloroanisole in various solvents.
Table 1. Quantum yields of the products formed by irradiation of 4-
chloroanisole.[a]
C
F1 +
Solvent
Fprod
MeCN
2: 0.003
0
MeCN/H2O (5:1)
MeOH
2: 0.007, 3a: 0.003
2: 0.08, 3b: 0.02
2: 0.038 (+0.012),[b] 3c: 0.02[c]
5: 0.07
0.22[a]
TFE
MeCN+DMB[d] (0.5m)
[a] [1]=1”10À3 m. [b] 4-Fluoroanisole. [c] A trace of product 4 was also
formed (the yield increases with a higher starting concentration, see
Table 2). [d] 2,3-Dimethyl-2-butene.
Scheme 1. Trapping of the phenylium ion by carbon nucleophiles.
further cationic intermediates, for example, the phenonium
ion (the cyclopropanespirocyclohexadienyl cation, II) in the
case of alkenes.[10] This latter species is an intermediate in
the solvolysis of phenethyl derivatives,[12] but has mostly
been directly detected under superacid conditions.[13] As
mentioned above, with benzene, the benzenium ion III is ex-
pected to be formed by trapping the phenylium ion with
benzene.
If the overall picture invoked above is correct, then this
photolytic reaction lends itself to the detection and study
under neutral, mild conditions of various cations that have
been invoked as intermediates, but rarely detected.[14,15]
Herein, we report a study aimed at the spectroscopic and ki-
netic characterization of the above cations by flash photoly-
sis with the support of product and computational studies to
prove their structures and roles.
Irradiation of this compound in acetonitrile, methanol, a
mixture of MeCN/H2O (5:1, v/v), and 2,2,2-trifluoroethanol
(TFE) gave anisole (2) as the main photoproduct (the only
one in MeCN) as well as products that result from the sub-
stitution of the chloro atom by the nucleophilic solvent (see
Tables 1 and 2); these were 4-methoxyacetanilide (3a) in
MeCN/H2O, 1,4-dimethoxybenzene (3b) in methanol, and 4-
(2,2,2-trifluoroethoxy)anisole (3c) in TFE. In TFE, some 4-
fluoroanisole was also formed along with a concentration-
dependent amount of the “self-attack” product, dimethoxy-
biphenyl 4. The reaction quantum yields were measured in
these solvents and are reported in Table 1. The values
varied from 0.003 in MeCN to 0.08 in MeOH. Previously we
have reported that irradiation in cyclohexane caused clean
reduction to 2.[17] The reaction in C6H12 was strongly
quenched in air-equilibrated solutions, whereas photode-
composition in the polar solvents mentioned above occurred
to the same extent (within Æ5%) in air- and argon-saturat-
ed solutions.
1030
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1029 – 1039

Phenylium-Derived Ions in Solution
FULL PAPER
pulse energy.[19a] Thus, by taking
into account the fact that
Table 2. Products and yields [%] obtained by the irradiation of 4-chloroanisole (0.05m) in neat solvent and in
the presence of nucleophiles.
DMB[a]
(0.5m)
1-Hexene
(0.5m)
1-Hexyne
(0.5m)
Benzene
(1m)
Thiophene
(1m)
CÀ
(S2O82À)=0.5, F
(SO4 )=0.7,
Solvent
Neat
F
N
N
and e450 =1600mÀ1 cmÀ1
for
G
U
N
N
ACHTREUNG
MeCN
2: 60[b]
5: 49[b]
2: 5
8a: 54
2: 12
8a: 30
9: 15
2: 15
12: 30
2: 11
2: 25
13: 33
2: 26
2: trace
CÀ [19b]
SO4
,
an absorption coeffi-
+
G
MeCN/H2O (5:1)
2: 53
3a: 22
2: 10
5: 21
2: trace
14: 65
14’: 15
was determined.[20]
12: 53
13: 53
Measurements on a solution
of 1 (1”10À3 m) in MeOH gave
values of DA at 460 nm (mea-
sured after completion of the
growing-in process) that were
found to depend linearly on the
laser pulse energy, which shows
that a monophotonic process
was occurring. By using the e460
value determined above and by
assuming that this does not
10: 14
2: 13
8a: 3
8b: 42
11b: 8
2: 4
8a: 36
8c: 10
MeOH
TFE
2: 63
2: 22
5: 16
2: 10
3b: 3
12: 17
2: 23
13: 15
2: 2
14: 14
14’: 4
3b: 29[b]
6a: 17
6b: 22[b]
2: 12
[c]
2: 24 (8)
3c: 15
2: 6
3c: 3
12: 60
2: 6
13: 61
2: trace
14: 52
14’: 5
5: 21
4: 20[b]
6a: 14
6c: 17
7c: 8[b]
11c: 40
[a] 2,3-Dimethyl-2-butene. [b] See ref. [17]. [c] 4-Fluoroanisole.
depend markedly on the sol-
+
C
vent, quantum yields for the one-photon formation of 1
Study of the intermediates: Laser flash photolysis at 266 nm
of argon-saturated solutions of 1 (ca. 1–1.5”10À3 m) was per-
formed in several solvents (measurements below 350–
360 nm at short delays after the pulse were perturbed by the
fluorescence of the solute). In cyclohexane an intense band
with lmax =335 nm (tailing into the visible region) and a life-
time of t=220 ns was observed after a delay of 40 ns after
the end of the pulse (Figure 1A). The transient was
were determined (see Table 1). In methanol, the value of F-
(1 ) was 0.08 at 3”10À5 m and reached a limit of around
0.22 at 1”10À2 m. The corresponding F0/F versus [1]À1 plot
is shown in the inset of Figure 2C.
+
C
Trapping experiments: The photochemistry of compound 1
was studied in the presence of various p nucleophiles, which
include alkenes (a mono- and a tetra-substituted derivative),
an alkyne, an aromatic, and a heteroaromatic derivative.
Product studies were carried out in the same organic sol-
vents as above with 1m additive. Addition of 2,3-dimethyl-
butene (DMB) in cyclohexane led to the formation of some
alkylated products along with 2, but the differences were
more marked in the other solvents, with trapping products
being formed with all of the additives tested, often as the
main products (see Scheme 3 and Table 2). In the case of
DMB, the allylated product 4-(1,1,2-trimethyl-2-propenyl)
anisole (5) was isolated as the only product in MeCN and
along with 2 in MeCN/H2O (5:1, v/v). In alcohols, the prod-
ucts 2 and 5 were accompanied by alkylanisole 6a, as well
as by the phenethyl ethers 6b and 6c in MeOH and TFE,
respectively, and in TFE by the benzyl ether 7c. In the pres-
ence of 1-hexene, photolysis in MeCN gave 4-(2-chlorohex-
yl)anisole (8a) in a yield of 54%. In hydroxylated solvents,
ethers were again formed along with 2, that is, 4-(2-meth-
oxyhexyl)- (8b) and 4-[2-(2,2,2-trifluoroethoxy)hexyl]anisole
(8c), but also the isomeric 1-methoxyhexyl (11b) and 1-
(2,2,2-trifluoroethoxy)hexyl (11c) derivatives in MeOH and
TFE, respectively. In the mixed solvent, product 8a re-
mained the main product, but was accompanied by alkene 9
and ketone 10, the latter probably resulting from the oxida-
tion of the corresponding benzylic alcohol. Further, we
tested an alkyne, namely, 1-hexyne, which produced 4-
(hexyn-1-yl)anisole (12) as the main product. With aromatic
derivatives, benzene gave biphenyl 13 as the main photo-
product and thiophene formed the 2-aryl- (14, main prod-
uct) and 3-arylthiophene (14’) derivatives.
quenched by oxygen with a rate constant kO =3”109 mÀ1 sÀ1
2
3
and was assigned to the triplet 1.[18]
The behavior was different in polar solvents. The end-of-
pulse difference spectrum recorded in MeCN showed an ab-
sorption tail in the region 350–400 nm (Figure 1B) that was
attributed to the triplet based on the clear-cut peak at
350 nm at 25 ns delay. The peak disappeared with a time
constant of 25 ns and a band with lmax =475 nm could be dis-
tinguished at a delay of 150 ns. In MeOH the initial triplet
absorption, which tailed from 360 nm into the visible region,
evolved with a time constant of tꢀ30 ns into a band with
lmax ꢀ440 nm (see the spectrum at 70 ns delay in Figure 2A),
which further shifted to 460–470 nm within 1 ms (see the
spectra at 900 ns in Figure 2A and the profile at 470 nm
shown in the inset). The formation kinetics of the 460–
470 nm band depended on the concentration of 1, in accord-
ance with a rate constant of 5”109 mÀ1 sÀ1. Finally, in both
MeCN/H2O (5:1, v/v) and TFE an end-of-pulse absorption
peak at 400 nm was observed, which evolved with a time
constant of around 10 and 22 ns into spectra with lmax =470
and 420 nm, respectively (Figure 3A and Figure 4A).
In an attempt to identify one of the transients, one-elec-
CÀ
tron oxidation of 1 by sulfate radicals (SO4 obtained from
2À
S2O8 by excitation at 266 nm) was carried out in an aque-
ous solution. The transient formed (with an absorption max-
imum at 460 nm) was virtually coincident with that obtained
from the photolysis of 1 in methanol at 900 ns, which was
+
C
thus identified as the radical cation 1 (Figure 2C). Oxida-
tion of 1 was assumed to be stoichiometric and gave a tran-
sient absorbance at 460 nm that increased linearly with
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1031

S. Monti, A. Albini et al.
Figure 2. Difference absorption spectra of a solution of 1 (1.3”10À3 m) in
argon-saturated MeOH at various delays from the end of a 266 nm laser
pulse (4.0 mJ per pulse). Insets: zero time at pulse onset. A) No addi-
tives. Inset: (DA at 360 (black) and 470 nm (gray). B) In the presence of
DMB (0.2m). Inset: DA at 380 (black) and 470 nm (gray). C) Solution of
Figure 1. Difference absorption spectra of a solution of 1 (1.3”10À3 m) in
argon-saturated C6H12 (A), MeCN (B), MeCN in the presence of DMB
(0.45m) (C) at various delays from the end of a 266 nm laser pulse
(4.0 mJ per pulse). Insets: zero time at pulse onset. A) (DA at 370 nm.
B) DA at 370 (black) and 470 nm (gray). C) DA at 370 (black) and
470 nm (gray).
1 (1”10À3 m) in argon-saturated MeOH at 1 ms ( ); aqueous solution of 1
*
2À
(5”10À4 m) in the presence of S2O8 (0.1m, ); the two spectra are nor-
&
malized at 460 nm. Inset: plot of F0/F versus [1]À1
.
The fluorescence of 1 was found to be unaffected by the
above additives (Stern–Volmer constant (KSV)<0.1mÀ1),
which indicates that the latter did not interact with the excit-
ed singlet state of the haloaromatic. Furthermore, it is im-
portant to note that, although the reactions reported above
were carried out in argon-saturated solutions, essentially the
same results were obtained when this step was omitted.
Laser flash photolysis experiments were also performed in
the presence of additives. In particular, the effect of DMB
1032
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1029 – 1039

Phenylium-Derived Ions in Solution
FULL PAPER
band with
a
maximum at
340 nm. In the presence of 1-
hexyne (Figure 4D) the end-of
pulse absorption at 400–440 nm
continued to grow together
with the band at 340 nm and
with a similar time constant.
The formation of a new absorp-
tion band that peaked at 370–
380 nm was observed with ben-
zene (5”10À2 m), which was
concomitant with the decay of
the initial absorption tail above
400 nm (time constant ca. 20 ns,
Figure 4E). In the presence of
thiophene (Figure 4F), an ab-
sorption band with a maximum
at 360 nm and a tail in the visi-
ble region formed with a time
constant of around 13 ns in the
Scheme 3. Photochemical reaction of 4-chloroanisole in the presence of var-
ious p nucleophiles.
was explored in various solvents. In cyclohexane it was
found that DMB quenched the 335 nm transient (kDMB ꢀ8”
107 sÀ1 mÀ1). In more polar media like MeCN, MeOH, and
MeCN/H2O (5:1, v/v), DMB had little effect on the end-of-
pulse spectra, but prevented the formation of the transient
absorbing around 460 nm that is otherwise observed in these
solvents (see Figures 1C, 2B, and 3B).
More informative results were obtained in TFE in which
the clear-cut band present in neat solvent at 400 nm was no
longer observed with 0.2m DMB and was replaced by a
band at 440 nm, which had completely formed 10 ns after
the pulse (Figure 4B). Concomitant with the decay of the
band at 440 nm, a sharp and intense absorption grew at
320 nm in an apparently first-order process (kꢀ5”107 sÀ1)
with an isosbestic point at 390 nm. With 0.04m DMB, the
spectrum at 10 ns had a band at lmax =420 nm, but subse-
quent absorbance changes and kinetics were similar. These
results support the proposal that the initial transient at
400 nm (decay rate ꢀ108 sÀ1) was partially quenched by
DMB and its residual absorption at 10 ns added to the parti-
ally formed intermediate observed at 440 nm (decay rate
ꢀ5”107 sÀ1). Further tests were performed at other DMB
concentrations and the double reciprocal plot of the absorb-
ance at 320 nm versus [DMB] was found to be linear (see
Figure 5).
Figure 3. Difference absorption spectra of a solution of 1 (1.3”10À3 m) in
argon-saturated MeCN/H2O (5:1, v/v) at various delays from the end of a
266 nm laser pulse (4.0 mJ per pulse). Insets: zero time at pulse onset.
A) No additives. Inset: DA at 410 (black) and 470 nm (gray). B) In the
presence of DMB (0.2m). Inset: (DA at 440 nm.
Similar conditions were chosen to explore the effect of
other additives (0.2m). In the presence of 1-hexene (Fig-
ure 4C) the end-of-pulse absorption band at around 440 nm
disappeared (with a time constant of ca. 10 ns) to give a
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1033

S. Monti, A. Albini et al.
Figure 4. Difference absorption spectra from a solution of 1 (1.3”10À3 m) in argon-saturated TFE with various additives at various delays from the end of
a 266 nm laser pulse (4.0 mJ per pulse). Insets: zero time at pulse onset. A) No additives. Inset: (DA at 410 nm. B) In the presence of DMB (0.2m).
Inset: (DA at 440 nm. C) In the presence of 1-hexene (0.2m). Inset: DA at 440 nm. D) In the presence of 1-hexyne (0.2m). Inset: (DA at 440 nm. E) In
the presence of benzene (0.045m). Inset: (DA at 380 (black) and 440 nm (gray). F) Solution of 1 (1.8”10À3 m) in the presence of thiophene (0.062m).
Inset: DA at 450 (black) and 480 nm (gray).
presence of additive (6”10À2 m). No concomitant decrease
of the absorption in any other region of the spectrum was
observed.
Information). The potential energy surfaces (PESs) for the
reactions of both cations with ethylene, acetylene, and ben-
zene were calculated at the same level of theory and the
UV spectra of the putative intermediates of the trapping
processes were determined on the minima obtained by a
time-dependent density functional theory (TD-DFT)
method. The spectra are shown in Figure 6A–D (a fixed
FWHM of 3000 cmÀ1 was assumed, see the Supporting Infor-
mation) and comparisons between the experimental findings
Calculations: The structures and energies of the singlet
(115+) and triplet (315+) cations (see Scheme 4) were calcu-
lated at the UB3LYP 6-31G(d) level of theory. It was found
that the singlet was stabilized by a mere 0.7 kcalmolÀ1 (in
MeCN bulk) with respect to the triplet (see the Supporting
1034
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1029 – 1039

Phenylium-Derived Ions in Solution
FULL PAPER
electrons in the molecular plane, the other one delocalized
in the p system) with the charge distributed over the ring.
The triplet ions formed a slightly stabilized complex, but did
not add to n nucleophiles (e.g., H2O), whereas addition to p
nucleophiles was predicted to occur via a diradical cation in-
termediate. The purpose of this study was to support this
mechanism by detecting the relevant intermediates and es-
tablishing a viable kinetic scheme. For the processes dis-
cussed in the following see Scheme 4.
There is no doubt that the transient observed in cyclohex-
3
ane has to be attributed to triplet 1 on the basis of previous
reports in the literature, the known efficient intersystem
[18,21]
crossing (ISC)
in anisoles, and the observed oxygen
quenching. The only reaction to take place under these con-
ditions was homolysis, which finally yielded 2 via the phenyl
radical. This state was liable to moderate quenching by
DMB (kꢀ8”107 sÀ1 mÀ1), to form an exciplex. Under these
conditions, a mixture of reduced 2 and alkylated products
was formed, as previously determined.[17]
Figure 5. Double reciprocal plot of the maximum absorbance at 320 nm
versus the concentration of DMB after exposing a solution of 1 (1.0”
10À3 ) in argon-saturated TFE to a 266 nm laser pulse (4.0 mJ per pulse).
and the features of the PES are discussed below. The calcu-
+
C
lated spectrum of radical cation 1 is shown in Figure 6E.
In the other solvents tested, the transient absorption in
the UV/Vis spectrum at the end of the pulse was much
weaker and shorter-lived (rate of disappearanceꢁ3”
107 sÀ1). In contrast, the fluorescence intensity did not
change by more than 10% upon varying the solvent. This
fact, together with the evidence that will be discussed below,
such as the trapping by p nucleophiles and the minimal
effect of oxygen on the photoreaction, supports the proposal
Mechanism: A general mechanistic frame for the photo-
chemical reactivity of phenyl halides with a strong electron-
donating substituent in the para position has previously
been presented that is based on both experiments and on
DFTand CASSCF calculations of the 4-amino derivative. [10]
Thus, in polar solvents, heterolysis of the triplet state of 4-
chloroaniline led to the triplet cation 4-NH2-C6H4+, which is
the lowest-energy spin state of this derivative. Contrary to
the singlet cation localized at C1 (p6s0 structure), which re-
acted unselectively with both n (e.g., water) and p nucleo-
philes (e.g., alkenes), the triplet cation had a p5s1 structure;
this gave C1 a triplet carbene character (one of the unpaired
3
that in polar solvents cleavage proceeded again from 1, but
by a faster process, namely, heterolysis, to yield the triplet
3
phenylium ion 15+. In place of the 335 nm absorption, in
TFE a different short-lived absorption that peaked at
400 nm was detected at the pulse-end. The same transient
could also be distinguished in MeCN/H2O (5:1, v/v), al-
Figure 6. TD-UB3LYP 6-31G(d) calculated spectra for A) 4-methoxyphenyl cation 315+ (115+ has a negligible absorption in this region), B) intermedi-
+
+
+
ates 16+ (long l) and 17+ (short l), C) intermediates 19’+ and 19 , D) intermediate 20 , and E) radical cation 1
.
C
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1035

S. Monti, A. Albini et al.
Scheme 4. Overall mechanism for the photochemical reaction of 4-chloroanisole.
+
C
Radical cation 1 was apparent in MeOH because this sol-
though it was partially superimposed on another signal (at
460–470 nm, Figure 3A) and was likely to contribute to the
absorbance changes observed at 70 ns in MeOH (Figure 2A,
see below). This band was thus attributed to the triplet
C
vent rapidly reduces 15 to anisole (2) and prevents back-
electron transfer [reverse of Eq. (1)]. However, this is prob-
ably a general mechanism, although the amount of inter-
mediates accumulated depended on the medium. The results
in MeCN/H2O were intermediate between those obtained in
TFE and MeOH, as one may expect from the intermediate
3
cation and indeed fitted well the calculated spectrum of 15+
1
(lmax
A
had no significant absorption in the near-UV region (see the
Supporting Information). The lifetime of the transient (t
ꢀ20 ns) was appropriate for a cation in a non-nucleophilic,
ion-stabilizing solvent, such as TFE (relevant solvent param-
eter Y=1.04, to be compared with Y=À0.5 evaluated for
MeCN/H2O (5:1, v/v) and Y=À1.09 for MeOH).[22]
value of Y, with both 15+ and 1 detectable, although with
less intense signals than in TFE and MeOH, respectively.
This is not surprising, as in poorly ion-stabilizing MeCN
(Y=À3.45), the quantum yield of the reaction was low and
the transients almost undetected.
3
+
C
In TFE the free cation was formed and reacted within
around 50 ns. In a less ion-stabilizing solvent, such as metha-
nol, cage escape and reaction occurred on a somewhat
Most of the end-products observed were those expected
from cation 15+ by reaction with the starting material or
3
with the solvent. The reaction with 1 gave either reduced 2
(by electron transfer) or self-attack diphenyldiamine 4 (by
addition, see Scheme 4). Competition between the two pro-
cesses (ket/ksa) depended on the medium and addition was
important in an ion-stabilizing, less nucleophilic solvent,
such as TFE (see Table 2). Reactions with the solvent in-
volve fluoride abstraction from TFE (a process previously
observed on decomposition of benzenediazonium salts in
this solvent)[23] and solvolysis. The latter reaction was that
3
longer time scale (ca. 300 ns) with the band of triplet 15+
less clearly discerned in the initial phase and that of the rad-
+
C
ical cation 1 (maximum at 460–470 nm, as proven by its
CÀ
generation by SO4 oxidation and calculations) emerging
later (see inset of Figure 2A). This revealed (one of) the re-
duction mechanism(s), namely, the electron-transfer process
[Eq. (1), Scheme 4].
315þ þ 1 ! 15C þ 1Cþ
expected of singlet 15+ and gave products 3a–c. However,
these were not formed in the presence of p nucleophiles;
1
ð1Þ
The reaction rate depended on the starting concentration
arylated products, diagnostic of the triplet cation, formed in-
3
of 1 with a bimolecular rate constant of ketffi5”109 mÀ1 sÀ1.
stead. Clearly, 15+ was generated as the first intermediate
1036
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1029 – 1039

Phenylium-Derived Ions in Solution
FULL PAPER
3
by photoheterolysis of 1, but in the absence of a convenient
from 1-hexene by a hydride shift). Such a rearrangement
was expected and is diagnostic of a carbocation, but again
had been previously reported to occur from the phenonium
ion only under strongly acidic conditions.[13]
trap, intersystem crossing to the almost isoenergetic singlet
(calculated DGST =0.7 kcalmolÀ1) occurred. Singlet 115+ is
an unselective electrophile and gave ethers 3b and 3c with
MeOH and TFE, respectively. In MeCN/H2O addition to
the solvent gave 3a [Ritter reaction, see Eq. (2)].
3
Summing up, DMB efficiently quenched 15+ to form, as
3
predicted by calculations, diradical adduct 16+ and closed-
shell 17+, and the course of the reaction could be followed
step-by-step by flash photolysis. A similar course was fol-
lowed with 1-hexene, with the first intermediate barely dis-
tinguishable by a decreasing absorbance at 440 nm, but a
Arþ þ MeCꢂN ! MeCþ¼NÀAr ! MeCð¼OÞNHAr
ð2Þ
The formation of such products allowed us to assess the
singlet reaction path, even though 115+ did not absorb in
the accessible l window, and their absence when triplet di-
agnostic products were formed allowed the chemistry of the
two spin states to be distinguished.
Trapping by p nucleophiles was best followed in TFE in
which the formation of the phenylium ion was conspicuous.
Calculation of the PES for the reaction of the triplet cation
with ethylene supported the proposal that addition to al-
kenes led to a single-bonded triplet diradical as the first in-
strongly absorbing phenonium ion was observed (17+, lmax
=
340 nm, Figure 4C).
With alkynes, calculation of the PES predicted the forma-
3
tion of an open-chain triplet cation 19’+ as the first inter-
mediate, which then collapsed (upon ISC) to the ring-closed
vinylenephenonium 19+.[15a,b,24] In the experiment with 1-
hexyne a conspicuous transient was detected (lmax =340 nm,
Figure 4D) although the final spectrum retained a shoulder
at 400–440 nm. The observed transient corresponded to that
of 319’+ (lmax
less easily observed because of the blueshifted spectrum
(lmax(calcd)=262 nm, see Figure 4D and Figure 6C). Reduc-
A
À
termediate. Formation of a second C C bond and intersys-
tem crossing led to the strongly stabilized (singlet) phenoni-
um ion.[10] This mechanism fitted nicely with the sequence of
events observed in the flash photolysis experiments per-
formed in the presence of DMB (0.2m). Thus, complete
AHCTREUNG
tion of these high-energy ions to give 12 as the end-product
is in accordance with previous findings.[25]
With benzene as the trap, an open-chain triplet adduct
was not markedly stabilized and a single transient (strong
absorption with lmax =380 nm) could be observed by flash
photolysis. This was attributed with confidence to benzeni-
um cation 20+ on the basis of its closeness to the calculated
band at 385 nm (see Figure 6D). This transient absorption
was identical to that observed by Steenken, McClelland and
co-workers by photolysis of phenyldiazonium tetrafluorobo-
rate in the presence of triisopropylbenzene.[7] Note, howev-
er, that photolysis of the diazonium salt yielded the singlet
cation as a nonselective intermediate. Indeed, these authors
were able to observe cation 20+ only in an extremely non-
nucleophilic solvent, such as 1,1,1,3,3,3-hexafluoroisopropyl
alcohol.[26] In our case, photolysis of 4-chloroanisole yielded
quenching of 15+ within the laser pulse was accompanied
3
by the appearance of a new transient (lmax =440 nm), which
in turn converted (kꢀ5”107 sÀ1) into a longer-lived species
(lmax =320 nm). These intermediates were identified as the
open-chain cation 16+ and phenonium ion 17+, respective-
3
ly, and indeed the experimental spectra closely correspond-
ed to those calculated for these species (lmax(calcd)=414
A
and 272 nm, respectively, see Figures 4B and 6B). Also, the
rate constant measured for the formation of 17+ is consis-
tent with the occurrence of a spin-forbidden process. The
trapping of triplet 15+ by DMB was efficient and the yield
3
of adduct cation 17+ reached a plateau at a DMB concen-
tration of ꢀ0.1m. The double reciprocal plot in Figure 5
shows a linear fit in agreement with Equation (3):
3
triplet cation 15+, which reacted preferentially with p nu-
cleophiles. The cationic adduct with benzene (at the same
concentration) was clearly detected in TFE and also in the
other polar solvents tested, although it was less clear-cut in
some cases. A heteroaromatic molecule, such as thiophene,
also gave a transient at lmax =360 nm, which was attributed
to the analogous cation 21+.
ꢀ
ꢁ
1
1
e320
kd
¼
1 þ
ð3Þ
e320½17þꢃ
kad½DMBꢃ
h
in which e320 is the molar absorption coefficient (optical path
1 cm), h is the overall yield of 17+, kd is the rate constant
for decay of the phenylium cation in the absence of trap,
and kad is the bimolecular rate constant for the addition re-
action. From the intercept/slope ratio, the value kad/kd =
120mÀ1 was extracted, which indicates that quenching of
315+ by DMB occurs with a rate constant close to diffusion
control (kadffi6”109 mÀ1 sÀ1), on the basis that kdffi5”107 sÀ1.
In turn, the final products were obtained from adduct cation
17+ either by deprotonation or by nucleophile addition, al-
though with the additional complication that a Wagner–
Meerwein shift of a methyl group (R’) to form a benzylic
cation (18+) might precede the final step (as shown by the
formation of product 7 and analogously the formation of 11
The overall mechanism proposed for the photochemical
reactions of 1 is presented in Scheme 4.
Conclusion
Conditions for observing a phenylium ion in solution have
been found for the first time, the role of this species in reac-
tions has been demonstrated, and the assignment is support-
ed by calculations. The agreement between predicted and
observed chemistry, as well as between the calculated and
observed spectra of the intermediates, has demonstrated the
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1037

S. Monti, A. Albini et al.
spin-dependent chemistry of the phenylium ion first suggest-
ed by Schuster and co-workers in 1995.[2f] In accordance
with previous reports,[6] we have shown by calculation that
singlet phenylium has no convenient absorption window and
may escape detection, but calculations and product studies
have demonstrated that it does play a role to play. In con-
trast to the singlet state, triplet phenylium is a synthetically
useful intermediate owing to its selective reactions with p
nucleophiles, which have been characterized in detail. As an
example, addition to an alkene is calculated to yield an
open-chain cation adduct that then converts by intersystem
Acknowledgements
We thank Dr. Francesco Manoli for technical support and the MIUR for
partial support of this work.
[1] a) P. J. Stang in Dicoordinated Carbocations (Eds.: Z. Rappoport,
P. J. Stang), Wiley, New York, 1997, p. 451; b) M. Hanack, L. R. Sub-
ramanian in Methoden Org. Chemie,Vol. E19C,Carbokationen,Car-
bokation-Radicale (Ed.: M. Hanack), Thieme, Stuttgart, 1990,
p. 249; c) M. Speranza, in Dicoordinated Carbocations (Eds.: Z.
Rappoport, P. J. Stang), Wiley, New York, 1997, p. 157.
[2] a) L. S. Romsted, J. Zhang, L. Zhang, J. Am. Chem. Soc. 1998, 120,
10046–10054; b) A. Chauduri, J. A. Loughlin, L. S. Romsted, J. Yao,
J. Am. Chem. Soc. 1993, 115, 8351–8361; c) C. G. Swain, J. E.
Sheats, K. G. Harbison, J. Am. Chem. Soc. 1975, 97, 783–790;
d) R. G. Bergstrom, R. G. M. Landells, G. W. Wahl, H. Zollinger, J.
Am. Chem. Soc. 1976, 98, 3301–3305; e) H. Zollinger, Diazochemis-
try I, VCH, New York, 1995; f) S. M. Gasper, C. Devadoss, G. B.
Schuster, J. Am. Chem. Soc. 1995, 117, 5206–5211; g) S. Milanesi,
M. Fagnoni, A. Albini, J. Org. Chem. 2005, 70, 603–610.
crossing into the phenonium cation (315+! 16+! 17+) and
in fact three spectra (very similar to those calculated) can
be observed that have an isosbestic point (Figure 4B) and
rate constants that fit expectation. Under such conditions,
none of the products expected from the singlet is observed,
thus excluding its formation as the primary intermediate
upon dechlorination, although they are found in the absence
of traps, which indicates that ISC to the singlet can occur as
a secondary step.
3
1
[3] a) Y. Himeshima, H. Kobayashi, T. Sonoda, J. Am. Chem. Soc. 1985,
107, 5286–5288; b) Y. Apeloig, D. Arad, J. Am. Chem. Soc. 1985,
107, 5285–5286; c) L. R. Subramanian, M. Hanack, L. W. Chang, M.
A Imhoff, P. v. R. Schleyer, F. Effenberger, W. Kurtz, P. J. Stang,
T. E. Dueber, J. Org. Chem. 1976, 41, 4099–4103; d) K. Laali, I.
Szele, K. Yoshida, Helv. Chim. Acta 1983, 66, 1710–1720.
Photoheterolysis of phenyl halides in the presence of suit-
able traps under neutral, mild conditions, has thus allowed
À
us to obtain cationic adducts by C C bond formation,
[4] C. Galli, Chem. Rev. 1988, 88, 765–792.
namely, phenonium ions with alkenes, phenylvinyl cations
(or vinylenephenonium ions) with alkynes, and a benzenium
ion from benzene. The ions have been identified under the
same (and very mild) conditions and full agreement between
experimental data and calculations was observed. These in-
termediates have been the object of intense investigation
and in some cases have previously been characterized, al-
though generally under acidic conditions. The development
of a mild method for their generation greatly enlarges the
choice of conditions available for their exploitation, and
thus, the control over the final outcome, which hopefully
will allow the use of these versatile intermediates for prepa-
rative purposes.[10]
[5] A. Filippi, G. Lilla, G. Occhiucci, C. Sparapani, O. Ursini, M. Sper-
anza, J. Org. Chem. 1995, 60, 1250–1264.
[6] D. M. Smith, Z. B. Maksic, H. Maskill, J. Chem. Soc.,Perkin Trans.
2 2002, 906–913.
[7] S. Steenken, M. Askokkuna, P. Maruthamuthu, R. A. McClelland, J.
Am. Chem. Soc. 1998, 120, 11925–11931.
[8] a) H. B. Ambroz, T. J. Kemp, Chem. Soc. Rev. 1979, 8, 353–366;
b) H. B. Ambroz, T. J. Kemp, G. K. Prybytniak, J. Photochem. Pho-
tobiol. A 1997, 108, 149–153.
[9] a) M. Winkler, W. Sander, Angew. Chem. 2000, 112, 2091–2094;
Angew. Chem. Int. Ed. 2000, 39, 2014–2016; b) M. Winkler, M.
Sander, J. Org. Chem. 2006, 71, 6357–6367.
[10] a) B. Guizzardi, M. Mella, M. Fagnoni, M. Freccero, A. Albini, J.
Org. Chem. 2001, 66, 6353–6363; b) M. Mella, P. Coppo, B. Guizzar-
di, M. Fagnoni, M. Freccero, J. Org. Chem. 2001, 66, 6344–6352;
c) M. Fagnoni, A. Albini, Acc. Chem. Res. 2005, 38, 713–721.
[11] M. De Carolis, S. Protti, M. Fagnoni, A. Albini, Angew. Chem. 2005,
117, 1258–1262; Angew. Chem. Int. Ed. 2005, 44, 1232–1236.
[12] a) D. J. Cram, J. Am. Chem. Soc. 1949, 71, 3863–3870; b) D. J.
Cram, J. Am. Chem. Soc. 1949, 71, 3871–3875; c) D. C. Noyce, R. L.
Castenson, D. A. Meyers, J. Org. Chem. 1972, 37, 4222–4223; d) S.
Sieber, P. v. R. Schleyer, J. Gauss, J. Am. Chem. Soc. 1993, 115,
6987–6988; e) E. Del Rio, M. I. Menendez, R. Lopez, T. L. Sordo, J.
Phys. Chem. A 2000, 104, 5568–5571; f) E. Del Rio, M. I. Menendez,
R. Lopez, T. L. Sordo, J. Am. Chem. Soc. 2001, 123, 5064–5068;
g) S. Nagumo, M. Ono, Y. Kakimoto, T. Furukawa, T. Hisano, M.
Mizukami, N. Kawahara, H. Akita, J. Org. Chem. 2002, 67, 6618–
6622.
[13] a) G. A. Olah, R. J. Spear, D. A. Forsyth, J. Am. Chem. Soc. 1976,
98, 6284–6289; b) G. A. Olah, N. J. Head, G. Rasul, G. K. S. Pra-
kash, J. Am. Chem. Soc. 1995, 117, 875–882.
[14] In earlier experiments we had detected a transient that was attribut-
ed to a phenonium ion in two instances, see: a) I. Manet, S. Monti,
P. Bortolus, M. Fagnoni, A. Albini, Chem. Eur. J. 2005, 11, 4274–
4282; b) I. Manet, S. Monti, M. Fagnoni, S. Protti, A. Albini, Chem.
Eur. J. 2005, 11, 140–151.
Experimental Section
The photochemical reactions were performed by using nitrogen-purged
solutions in quartz tubes in a multilamp reactor fitted with six 15 W phos-
phor-coated lamps (maximum emission=310 nm) for irradiation. The
photoproducts were isolated by chromatography and compared with au-
thentic samples or identified by chemical and spectroscopic analysis.
Nanosecond laser flash photolysis experiments were carried out by using
a JK-Laser Nd-YAG laser operated at l=266 nm and pulse FWHM=
20 ns; the minimum response time of the detection system was around
2 ns (see the Supporting Information). The structures and energies of the
singlet and triplet 4-methoxyphenylium ions and the PES for the reac-
tions with ethylene, acetylene, and benzene were calculated at the
UB3LYP/6-31G(d) level of theory by using the Gaussian 03 package[27]
(see the Supporting Information). The spectra of transients were calculat-
ed by the TD-DFT method (see the Supporting Information).
[15] Related cations have previously been detected by laser flash photol-
ysis. Benzenium ions were detected by the addition of the phenyli-
um ion, photogenerated from benzenediazonium salts, to benzenes,
see ref. [7]. Phenylvinylium ions have been generated by protona-
tion of phenylacetylenes, see: a) R. A. McClelland, F. L. Cozens, S.
1038
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1029 – 1039

Phenylium-Derived Ions in Solution
FULL PAPER
Steenken, Tetrahedron Lett. 1990, 31, 2821–2824; b) M. Kotani, S.
Kobayashi, J. A. Chang, J. Phys. Org. Chem. 2002, 15, 863–868.
[16] a) G. Grabner, C. Richard, G. Kçhler, J. Am. Chem. Soc. 1994, 116,
11470–11480; b) A. P. Durand, R. G. Brown, D. Worrall, F. Wilkin-
son, J. Chem. Soc.,Perkin Trans. 2 1998, 365–370.
[24] a) K. Van Alem, G. Belder, G. Lodder, H. Zuilhof, J. Org. Chem.
2005, 70, 179–190; b) T. Okuyama, Acc. Chem. Res. 2002, 35, 12–
18; c) M. Kotani, S. Kobayashi, M. Mishima, Y. Hori, Chem. Lett.
2003, 32, 294–295; d) F. L. Cozens, V. M. Kanagasabapathy, R. A.
McClelland, S. Steenken, Can. J. Chem. 1999, 77, 2069–2082.
[25] S. Protti, M. Fagnoni, A. Albini, Angew. Chem. 2005, 117, 5821–
5824; Angew. Chem. Int. Ed. 2005, 44, 5675–5678.
[17] S. Protti, M. Mella, M. Fagnoni, A. Albini, J. Org. Chem. 2004, 69,
3465–3473.
[18] a) H. Lemmetyinnen, J. Konijnenberg, J. Cornelisse, C. A. G. O.
Varma, J. Photochem. 1985, 30, 315–338; b) J. P. Da Silva, L. F.
Vieira Ferreira, I. Ferreira Machado, A. M. Da Silva, J. Photochem.
Photobiol. A 2006, 182, 88–92.
[26] The advantage of using hexafluoroisopropanol for stabilizing radical
cations has been extensively documented, see: L. Eberson, M. P.
Hartshorn, O. Persson, J. Chem. Soc.,Perkin Trans. 2 1995, 1735–
1744.
[19] a) However, the (much weaker) signals observed in MeCN in-
[27] Gaussian 03, Revision B.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
+
C
creased nonlinearly, which shows that in this case 1 arose in part
by two-photon ionization. In order to separate one- and two-photon
contributions, the absorbance data were treated by using the equa-
tion DA=aP+bP2, which is valid in the low-pulse energy region, in
which P is the pulse energy and a and b are factors characteristic of
the one- and two-photon processes, respectively. a is proportional to
the quantum yield of the one-photon production. It was found that
the one-photon process is negligible in MeCN; b) W. J. McElroy, J.
Phys. Chem. 1990, 94, 2435–2441.
[20] The lower intensity of the band at around 305 nm in the spectrum in
the presence of sulfate in Figure 2C is probably owing to the deple-
tion of S2O82À, which begins to absorb in this region.
[21] J. Den Heijer, O. B. Shadid, J. Cornelisse, E. Havinga, Tetrahedron
1977, 33, 779–786.
[22] a) S. Winstein, A. H. Fainberg, E. Grunwald, J. Am. Chem. Soc.
1957, 79, 4146–4155; b) S. Winstein, A. H. Fainberg, J. Am. Chem.
Soc. 1957, 79, 5937–5950.
[23] P. S. J. Canning, K. McCrudden, H. Maskill, B. Sexton, Chem.
Commun. 1998, 1971–1972.
Received: July 6, 2007
Published online: November 14, 2007
Chem. Eur. J. 2008, 14, 1029 – 1039
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1039